Assessing the validity of corneal power estimation using conventional keratometry for intraocular lens power calculation in eyes with Fuch’s dystrophy undergoing Descemet membrane endothelial keratoplasty

Intraocular lens (IOL) power calculation is traditionally based on keratometers that estimate the corneal refractive power from anterior corneal measurements, by using a standardized fictitious refractive index (1.3320) referring to a theoretical single refractive lens representing both corneal surfaces. This so-called conventional keratometry (K) is used in IOL Master PCI devices for corneal power estimation and assumes a constant posterior to anterior corneal curvature radii ratio (PA ratio or R_PA) [1]. This assumption leads to sufficient refractive outcomes when IOL power calculation is performed in virgin eyes [2, 3].

However, when anterior corneal radius is altered by corneal refractive surgery, the PA ratio is disrupted and the usual keratometric refractive index becomes invalid. As a consequence, after myopic correction, conventional K readings usually overestimate corneal power and the resulting IOL power is underestimated, leading to postoperative hyperopia [4].

Similar, in eyes with Fuch’s dystrophy, a hyperopic shift has been reported when cataract surgery is combined with a Descemet membrane endothelial keratoplasty (triple DMEK) [5‐13]. This has been attributed to regression of the posterior stroma edema associated with a steepening of the posterior corneal curvature [5‐13]. However, as the posterior corneal curvature is not measured directly, neither a preoperative flat posterior corneal curvature, nor a postoperative change of posterior corneal radius is considered in the IOL power calculations based on measurement of the anterior corneal surface and conventional K [1]. Therefore, the decisive ratio for IOL power calculation in eyes with Fuch’s dystrophy, is the ratio between postoperative posterior corneal radius, once stable refraction is achieved, and preoperative anterior corneal radius, when conventional K is performed (Table 1).

We hypothesized that when this postoperative posterior to preoperative anterior corneal curvature radii (PPPA) ratio differs from the PA ratio in virgin eyes, it will render the keratometer index invalid and cause a hyperopic shift after triple DMEK, similar to IOL power calculation in eyes after myopic photoablative refractive surgery [4].

The aim of this study was to assess the validity of conventional K for IOL power calculation in eyes with Fuch’s dystrophy undergoing triple DMEK using Scheimpflug imaging (Pentacam HR). Therefore, we compared the postoperative posterior to preoperative anterior corneal curvature radii (PPPA) ratio in eyes with Fuch’s dystrophy undergoing DMEK with the PA ratio calculated in healthy eyes.

Finally, to guide further IOL power calculation methods, the most appropriate ratio and keratometer index for the conversion of preoperative anterior corneal radius to postoperative posterior and total corneal powers was determined in a theoretical model.

A total of 125 eyes of 125 patients who underwent DMEK or triple DMEK and 125 eyes of 125 patients without corneal pathologies were included in this retrospective study. Patients’ characteristics are summarized in Table 1. Postoperative changes relative to baseline are shown in Table 2.

Eyes after DMEK surgery showed a statistically significant thinner central corneal thickness, steeper posterior corneal radius, and lower PA ratio when compared to eyes without corneal pathologies (Table 3).

The PPPA ratio and keratometer index determined in eyes with Fuch’s dystrophy undergoing DMEK was significantly different (P < 0.001) to the keratometer index and PA ratio in eyes without corneal pathologies (Table 3). In a post hoc power analysis, the calculated effect size was 0.62 (ratios) and 0.63 (keratometer indices). With an alpha of 0.05, this led to a power of 0.99 in both cases.

Using the geometric mean R_PA^Control (0.822) and n_c^Control (1.3283) to convert the anterior corneal radius before DMEK to power leads to significant (P < 0.001) underestimation of the negative postoperative posterior (∆D_P = − 0.14D ± 0.30) and thus overestimation of the total corneal power (∆D_Total = −0.45D ± 1.08) (Table 4).

The lowest prediction errors were found using the mean fictitious R_PA^FECD/DMEK (0.806) and n_c^FECD/DMEK (1.3273) to predict the postoperative posterior (∆D_P = − 0.01D ± 0.30) and total corneal powers (∆D_Total = − 0.32D ± 1.08) (Table 4).

Figures 2 and 3 show the Bland-Altman plots comparing predicted and measured postoperative posterior and total corneal powers (Figs. 2 and 3).

Ours is the first study to demonstrate that using conventional keratometry to estimate postoperative corneal refractive power in patients with FECD undergoing DMEK surgery, leads to an overestimation of corneal power. Through the introduction of a fictitious PPPA ratio and keratometer index, we have found a promising way to adjust conventional keratometry and provide the necessary correction.

The 4 potential sources of error in IOL calculation are corneal curvature measurement, axial length measurement, effective lens position estimation, and the IOL calculation formula [15‐17].

This study focused on corneal power measurement that accounts for approximately two-thirds of the total dioptric power of the eye [17]. If the calculation of corneal power is inaccurate, it will have profound consequences on subsequent steps in the calculation of IOL power [18‐21].

Optical biometers such as the IOL Master 500 (Carl Zeiss Meditec, Jena, Germany) do not measure the posterior cornea directly, but instead account for it using a fictitious refractive index of the cornea, under the assumption that the posterior to anterior corneal curvature radii ratio is constant in all eyes [1]. In this study, the posterior corneal power changed significantly after DMEK surgery, a finding in line with all related studies presented in the literature [5‐12] (Table 2).

From the combined information presented above, we conclude that the relationship between the postoperative posterior corneal radius—once stable refraction is achieved—and to the preoperative anterior corneal curvature—when conventional K is performed—is decisive for IOL power calculation (IOL Master 500) in eyes undergoing DMEK surgery (Fig. 1).

Furthermore, we have demonstrated that this ratio is significantly different to the PA ratio in healthy eyes, so that the keratometric index and the corneal power derived from it by this method will become invalid [18, 19].

This is similar to the situation with altered PA ratio after myopic laser vision correction (e.g., LASIK, PRK), where conventional K overestimates corneal power, which leads to a biased IOL calculation and hyperopic outcome after cataract surgery [4, 19, 22‐25].

To counteract this, modern optical biometers (e.g., IOLMaster 700) are capable of measuring both anterior and posterior corneal curvatures, to assess the total keratometry (TK). With the use of innovative IOL calculation methods, such as the Haigis-TK or Barret True-K formula, this enables a more precise outcome after cataract surgery with previous LASIK or PRK [4, 26, 27].

In this context, it is important to mention that in eyes with Fuch’s dystrophy undergoing DMEK, the posterior corneal curvature changes significantly after the performed surgery. Therefore, considering the preoperative posterior corneal curvature in IOL power calculation would lead to an overestimation of the postoperative corneal power (Table 4).

In our study, DMEK leads to a significant thinner central corneal thickness, steeper posterior corneal curvature, and thereby lower PPPA and postoperative PA ratios when compared to healthy eyes (Table 3). Similar results are presented in the literature. Arnalich-Montiel et al. showed a thinner cornea with a steeper pachymetric progression from the thinnest point to the periphery in eyes after DMEK compared to normal corneas [28].

This difference could be simply explained by the fact that DMEK surgery does not replace the entire endothelium and Descemet membrane (DMEK roll size usually lies between 8.0 and 9.25 mm).

However, a standard IOL power calculation using conventional K is inaccurate in the case of triple DMEK and will lead to a significant underestimation of posterior negative corneal power and hence overestimation of total corneal power (Table 4).

In the clinical practice, DMEK surgeons have noticed this problem and aim to achieve a more myopic postoperative outcome by choosing a refractive target of − 0.5 to − 1D to compensate for this error [8]. In the future, IOL power calculation could be optimized for eyes with FECD undergoing triple DMEK surgery by using adjusted conventional keratometry to improve the predictability of IOL power calculation and account for the postoperative hyperopic shift [18, 21, 29‐31].

In comparison to that, a higher postoperative hyperopic error has been reported in eyes undergoing Descemet stripping endothelial keratoplasty (DSEK) combined with cataract surgery [32]. This could also be explained by a non-physiologic postoperative posterior corneal curvature that is also overestimated by conventional K. Goldich et al. found a significantly steeper posterior corneal radius after DSEK compared to eyes undergoing DMEK. In contrast, there was no difference when comparing the anterior corneal curvature preoperative and postoperative in between the groups [33].

Using the PPPA ratio presented in this study (0.806) enhances the accuracy of the refractive prediction of postoperative posterior corneal power in comparison with the PA ratio of eyes with FECD (0.900) or healthy corneas (0.822) (Fig. 2).

Compared to that, using the fictitious keratometer index (1.3723) to preoperatively predict postoperative total corneal power leads to a minimal additional overestimation of that power on average. This is caused by a significant postoperative flattening of the anterior corneal curvature in the present data (Fig. 3).

This is similar to the results of our previous study, where we noted a more strongly correlation between preoperative corneal parameters (e.g., PA ratio, Asph. Q) and the change in refractive power of the posterior corneal surface compared to the change in total corneal power [14].

Regarding the change of the anterior corneal curvature after DMEK, different results are presented in the literature. Kwon et al. found no significant alteration of simulated keratometry after surgery [10], whereas van Dijk et al. showed an ongoing significant change in mean anterior corneal curvature [11].

The variation between predicted and measured postoperative total corneal power (± 1.08D) indicates that changes of anterior corneal curvature could explain the problem of varying postoperative refractive outcomes after triple DMEK that go from myopia to hyperopia.

However, a keratometer index cannot account for these changes [34] and the prediction of the individual refractive outcome in eyes with Fuch’s dystrophy prior to Triple DMEK will remain a challenging task for surgeons.

This pilot study was limited by a relatively small sample size and the retrospective design. This may have affected the results to a minimal extent. Nevertheless, a high power was achieved in the post hoc power analysis.

Moreover, the results presented in this monocenter study may be skewed by specific surgical details (incision at 12 h and 8.75-–9.00-mm transplant); this ratio and keratometer index could easily be evaluated for the different surgeons or different centers. In any case, multicenter studies with a higher number of patients are needed.

In conclusion, the postoperative posterior to preoperative anterior corneal curvature radii ratio is the decisive ratio for IOL power calculation in eyes undergoing triple DMEK and differs significantly from the PA ratio in healthy eyes. Conventional K for IOL power calculation in eyes with FECD undergoing triple DMEK surgery should therefore be corrected to avoid overestimation of postoperative corneal power and to minimize the risk of a hyperopic shift. The newly derived PPPA ratio and corresponding fictitious keratometer index n_c^FECD/DMEK represent a promising corrective tool. Further studies with innovative IOL power calculation methods are needed to evaluate the utility of this parameter in patients undergoing triple DMEK.